Wednesday, October 18, 2006

Mind the Lag

Something I noticed which I have not the time to explore, but perhaps one of you aspiring junior deputy economists can figure this one out.

But previous to the past two recessions we’ve had, unemployment peaked at the end of recessions. This only makes sense as so long as the economy is retracting, then unemployment would go down until growth took over again.

But now it seems unemployment lags behind the economy. Eyeballing it, it looks like about a year to a year and a half lag before unemployment peaks and then starts to recover;

5 comments:

Hydrick said...

I'm by no means an economist, so I'm probably dead wrong, but could this be related to outsourcing? The scenario I'm thinking of is that as the economy continues in a recession here, businesses outsource operations to reduce costs, increasing unemployment. This outsourcing continues until the economy is in an established growth pattern, at which point businesses begin hiring in the US again. I may be horribly wrong, but that was just the thought I had on it.

Gabriel M said...

How does the NBER date recessions? 2 consecutive quarters of negative growth, until 2 consecutive quarters of positive growth? --

You're right. Weird stuff!

Anonymous said...

If the unemployment numbers peak so long after the official figures declare that the market has recovered, you should question the veracity of the official statistician(s), and suspect that they have made an political assessment rather than an honest one.
There is a secondary possibility but it would be very dangerous indeed.

Anonymous said...

Just from looking at the chart, the nature of the movement of the unemployment rate is different in the past two recessions from those prior to the last two. In the latest two the unemployment rate had a drastic move upward, followed by a brief levelling off, and then a small spike up before falling again. The magnitude of the small spike seems to be close to a golden ratio(which in itself is somewhat interesting, although I can offer no explanation for that phenomenon), ie 38% of the total unemployment growth from trough to peak. Contrast this with the behavior in previous recessions, where the end of the recession marked the top of unemployment - in those, there was no levelling off of the increase of unemployment prior to the end of the recession.

Perhaps it has to do with human nature and the way unemployment statistics are calculated... For example, someone gets laid off at the end of the recession, is on unemployment for a while(which was extended by Congress in the last recession for sure, perhaps in the previous one as well). The economy has started to turn by the time they're off unemployment, but businesses are still tentative about the recovery and those who do have jobs are happy to have them and may be willing to do more work than they were accustomed to doing, in essence being more productive. In the meantime, some of those who were laid off at the end of the recession may have dropped off of the unemployment rate calculation denominator(total workforce) being considered as not looking for jobs, making the rate 'artificially' higher. It also appears that the unemployment rate has not declined as quickly in the past two recessions as it had in those prior, perhaps lending credence to this hypothesis, being that those who were not considered to be looking for work might revise their outlook and begin looking for work once again, adding their contribution back into the denominator.

It would be illustrative to have the raw data for the unemployment rate calculation to see the change in total workforce vs those being counted as unemployed. Productivity stats would also help find the cause - see whether productivity was still increasing after the past two recessions ended and to see whether it differed from those prior - I would guess that in the prior recessions productivity remained relatively flat while the unemployment rate dropped which would probably be different than in the last two, where I would guess productivity was still increasing.

A couple of other thoughts about the nature of the economy during the two time periods. One, our economy has definitely shifted to being more service-based in the past 20 years. And, two, there has been a lot more illegal immigration during the same time period and whether illegals are counted in unemployment statistics would definitely have an effect on the rate.

That's just my armchair take as an equity trader though...

Anonymous said...

Another possibility is that some companies hold onto employees for as long as they can. It costs money to recruit and train a new workforce.

As a result, when the economy starts to pick up, they don't need to hire new workers. For example, I worked in a woodshop for several years. When things got slow, they would send us home early (without pay) on Fridays.

Combine this with the fact that recoveries, especially in the early phases are rarely even. Certain regions, industries, even companies, start picking up first.

This can lead to a situation in which some companies are picking up work, but because they some slack in their internal workforce, they don't need to hire new workers immediately. They can even go to overtime while they are determining if the new orders are a blip, or a signal that the overall economy has turned around.
At the same time, other companies are still hurting. Their orders have not yeat picked up. Some of them are finally reaching the point where they have to let people go, or risk losing the entire company. For other companies, the new born recovery has come too late, and they do go out of business.

To repeat, for most companies, the initial new orders can be met with their existing workforce, so there is no need for them to start hiring yet. Other companies, the ones not getting new orders yet, are still shedding workers.